I respectfully disagree that zeroing out reviews is a key to burnout mainly because your review pile is dictated by how much lessons you do and reviews you fail. Therefore, your review pile is only as big as you make it, and I think we are both trying to make the point that having too much lessons leads to a review pile that becomes too daunting to zero out daily, and in that regard I heartily agree.
The reason I recommend that one clears out their review pile is because items are supposed to appear when the system thinks youāre about to forget it (at least Anki does with FSRS) so if you donāt review those items for that specific day, technically you should have forgotten them after that day. Of course, Wanikaniās system isnāt as robust as FSRS (or SuperMemo for that matter), and each person retains at different lengths, I think pushing back your review pile is a bigger instigator of burnout simply because by the time you get to the items you pushed back, youāve (probably) forgotten them and thus feel disheartened that you keep making mistakes.
However all of these are from my personal observations of how I like to deal with WaniKani, as weāre finding, every one has a unique approach. I tend to trend towards burnout more if I make too many mistakes, and would prefer to keep my retention high, even if that means clearing out my review pile daily.
As weāre finding here, your approach is certainly different to mine and thatās great, because we can offer different perspectives to people reading these posts looking for advice and offer different options to tackling the sheer volume of learning WaniKani offers.
TLDR:
Lessons and mistakes dictate your review pile, so limiting lessons (to a level you can handle) enables you to zero out your reviews daily and thus keep your retention high (less mistakes) which brings your review pile down.
For actual values, keeping your apprentice numbers to around 90 = about 60-80 ish reviews daily. Lowering that value brings your daily reviews even lower.
The main problem I see with needing to clear the Review pile, is when to add new lesson(s). Can at least some lessons be added regularly?
Exactly 1 per day might be an answer, but then, whatās the best way to get lower SRS (i.e. Apprentice N) reviewed? Perhaps either by clearing the Review every day, or with userscripts/apps, sorting by lower SRS first ā I am reluctant to make either of those a recommendation, especially for those back from a long break, or having a sea of leeches.
I think low Apprentice, and perhaps also low Guru, is good place to monitor ā starting to memorize or not.
In my experience, the answer to better accuracy are
Not too big batches
Sufficient repeats for new items. Probably shorter intervals than given by SRS. Probably even immediate repeats, like A-B-A-B-A-B-ā¦
Memory aids help sometimes, not limited to mnemonics. Imagining a good scenery / scenario is another one. I prefer looking up for cross references until satisfied.
About clearing the Review pile, and trusting the SRS for long term memory, WaniKani not using FSRS or any adaptive and widening intervals is one of the reasons for exaggerated review pile. (And not testing Burned every once in a while.) Another thing is I donāt think I need long-term memory for disjointed vocabularies, but rather just medium-term memory, enough to ascertain the importance of vocabularies in immersion or usage.
But then, the solution for medium-term memory building isnāt necessarily SRS as well.
I can definitely agree with the points here. I think it just boils down to the unfortunate fact that WaniKani uses fixed intervals and further has intervals stop at around 5-6 months of initially seeing the word. So we never really get to stretch the legs of long term retention.
Regarding when to add lessons, I used to use Lesson Lock before they added fixed lessons. But I still appreciate the feature of locking reviews until the apprentice items fall below a specified number. So I might recommend that and finding a nice number that makes zeroing out reviews doable.
However, I also agree with there not being a great importance on having a large vocabulary only from brute forcing SRS. While I personally prefer doing it that way (and I donāt really recommend it due to workload issues), at a certain point, most of your exposure should come from immersion anyway and any words that arenāt repeated in your immersion are basically not realistically relevant to you and therefore can spare a dictionary check whenever you do encounter it, if at all.